From mboxrd@z Thu Jan 1 00:00:00 1970 From: Shmulik Ladkani Subject: Re: [PATCH] net: ip_finish_output_gso: If skb_gso_network_seglen exceeds MTU, allow segmentation for gre tunneled skbs Date: Fri, 12 Aug 2016 08:18:00 +0300 Message-ID: <20160812081800.02a98e17@halley> References: <1470726261-16371-1-git-send-email-wenxu@ucloud.cn> <20160810.173511.968926810628735179.davem@davemloft.net> <20160811224132.7a17f0c3@halley> <80d116d7-e61c-1bbd-64bf-e3b1f809419b@ucloud.cn> Mime-Version: 1.0 Content-Type: text/plain; charset=US-ASCII Content-Transfer-Encoding: 7bit Cc: David Miller , kuznet@ms2.inr.ac.ru, jmorris@namei.org, kaber@trash.net, yoshfuji@linux-ipv6.org, netdev@vger.kernel.org, wenx05124561@163.com, hannes@stressinduktion.org To: wenxu Return-path: Received: from mail-wm0-f47.google.com ([74.125.82.47]:34278 "EHLO mail-wm0-f47.google.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1750855AbcHLFSK (ORCPT ); Fri, 12 Aug 2016 01:18:10 -0400 Received: by mail-wm0-f47.google.com with SMTP id q128so1308503wma.1 for ; Thu, 11 Aug 2016 22:18:09 -0700 (PDT) In-Reply-To: <80d116d7-e61c-1bbd-64bf-e3b1f809419b@ucloud.cn> Sender: netdev-owner@vger.kernel.org List-ID: Hi, On Fri, 12 Aug 2016 11:51:07 +0800 wenxu wrote: > > And in b8247f095edd, the condition skb_iif also should be removed. > given: > ovs-internal-dev bridge with ovs-gre > > There are the same problem which the skb from local. There's no need to remove the skb_iif criteria: For the local bridge port, we have control over its mtu. This allows setting an mtu value that takes into account the gre encapsulation performed by other member ports. By doing so, locally generated traffic (on br0) will have a proper gso_size. OTOH if packet arrives from an ingress member port (such as tap0) the packet's gso_size could have been set by another system - which is not always under our control. ( see discussion in http://www.spinics.net/lists/netdev/msg385085.html ) Regards, Shmulik